This is (un) fortunately a problem I'm intimately familiar with. In the
end Ive pretty much always solved it by writing a small utility or script.
1: You need to normalize breakouts across carriers, this means expanding
to the longest match, so in the previous example:
Number dialed: 44-20-7499-9000
Carrier A: 44 - 0.0025
Carrier B: 442 - 0.0045
Carrier C: 44207 - 0.0085
you end up with:
44: carrier A - 0.0025
442: carrier A - 0.0025, Carrier B - 0.0045
44207: carrier A - 0.0025, Carrier B - 0.0045, Carrier C - 0.0085
Great now your routing table is instead of 215k entries, 1.3M but its
comparable.
If you have a cost cap, before you do the next part, strip all routes
that exceed it. You don't want pricing for routes you'll never use
influencing your rates.
2: For each destination drop your lowest cost and use some combination
of your tolerable route depth pricing * some margin. You might also
consider a smarter algo like dropping lowest if more than std dev away
from avg of next X carriers etc. Basically you dont want your price
forced below carrier 2/3 by an abnormally low 1 who in the end will
never complete calls satisfactorily for you.
3: Now, you need to de-duplicate, removing all routes whose price is
identical to their parent route (route stripping the right-most digit,
if that doesnt exist, strip again until you hit base country code)
4: Finally, take your rate deck to your sales team and listen to them
tell you how they cannot sell it because its more expensive than
<fly_ny_night_telecom>.
There are lots of other ways to do this, but i pretty much always
implement some flavor of this process.
FYI, after expansion, if you have the means, its always worth adding a
step that scans for fictitious codes. Occasionally IRSF perpetrators
will inject bogus country sub-codes in the hopes of getting FAS traffic
from fraudsters.
Hope that helps.
-Ryan
On 6/4/2019 7:10 AM, Shripal Daphtary wrote:
Hey group,
I have a question that I have been struggling with for years and have
never come up with a good solution for. It revolves around
International Rate Deck creation, but i guess it could be for any
tariff. We have multiple carriers for International, however, i'm
trying out Thinq right now so we can use their LCR. Our other
carriers aren't very successful with Intl. Thinq's rate deck to me is
6 carriers for each prefix, making it around 215,000 lines. The
carrier(s) that have the lowest cost for each prefix varies, so i
can't turn off the most expensive three or something like that.
I was thinking of taking the least expensive 3 carriers and then
averaging them and creating my rate from that average and then only
allow Thinq to go 3 carriers deep. Does anyone have any experience
with this? Are there any best practices?
The second part of the question is how does one calculate the profit
margin? Let's say you wanted to make 35% for retail and 20% for
wholesale, but if you call UK landline, the cost is only 0.004. Your
rate would be 0.0054 for retail and 0.0048, which is nothing. We
have been doing something like If your cost is less than 0.03, then
increase by 35% or 20% or whatever. however, that doesn't always work
if the cost is super close to your target.
Does anyone have any hard and fast rules that they use when creating
decks? is there software that can help my puny brain think through this?
Thanks !
Shri
_______________________________________________
VoiceOps mailing list
[email protected]
https://puck.nether.net/mailman/listinfo/voiceops
_______________________________________________
VoiceOps mailing list
[email protected]
https://puck.nether.net/mailman/listinfo/voiceops
